Modelling the effect of active roots on soil organic matter turnover

The aim of this experiment was to study the effect of living roots on soil carbon metabolism at different decomposition stages during a long-term incu bation. Plant material labelled with C-14 and N-15 was incubated in two con trasting soils under controlled laboratory conditions, over two years. Half the samples were cropped with wheat (Triticum aestivum) 11 times in succes sion. At earing time the wheat was harvested, the roots were extracted from the soil and a new crop was started. Thus the soils were continuously occu pied by active root systems. The other half of the samples was maintained b are, without plants under the same conditions. Over the 2 years, pairs of c ropped and bare soils were analysed at eight sampling occasions (total-, pl ant debris-, and microbial biomass-C and -C-14). A five compartment (labile and recalcitrant plant residues, labile microbial metabolites, microbial b iomass and stabilised humified compounds) decomposition model was fitted to the labelled and soil native organic matter data of the bare and cropped s oils. Two different phases in the decomposition processes showed a differen t plant effect. (1) During the initial fast decomposition stage, labile C-1 4-material stimulated microbial activities and N immobilisation, increasing the C-14-microbial biomass. In the presence of living roots, competition b etween micro-organisms and plants for inorganic N weakly lowered the measur ed and predicted total-C-14 mineralisation and resulted in a lower plant pr oductivity compared to subsequent growths. (2) In contrast, beyond 3-6 mont hs, when the labile material was exhausted, during the slow decomposition s tage, the presence of living roots stimulated the mineralisation of the rec alcitrant plant residue-C-14 in the sandy soil and of the humified-C-14 in the clay soil. In the sandy soil, the presence of roots also substantially stimulated decomposition of old soil native humus compounds. During this sl ow decomposition stage, the measured and predicted plant induced decrease i n total-C-14 and -C was essentially explained by the predicted decrease in humus-C-14 and -C. The C-14-microbial biomass (MB) partly decayed or became inactive in the bare soils, whereas in the rooted soils, the labelled MB t urnover was accelerated: the MB-C-14 was replaced by unlabelled-C from C de rived from living roots. At the end of experiment, the MB-C in the cropped soils was 2.5-3 times higher than in the bare soils. To sustain this biomas s and activity, the model predicted a daily root derived C input (rhizodepo sition), amounting to 5.4 and 3.2% of the plant biomass-C or estimated at 4 6 and 41% of the daily net assimilated C (shoot + root + rhizodeposition C) in the clay and sandy soil, respectively.